Adversarially Learned Anomaly Detection

TL;DR

This paper introduces ALAD, a novel anomaly detection method using bi-directional GANs that leverages adversarially learned features and cycle-consistencies to improve detection accuracy and speed.

Contribution

The paper proposes ALAD, a bi-directional GAN-based anomaly detection framework that enhances performance and efficiency over previous GAN-based methods.

Findings

Achieves state-of-the-art results on multiple datasets

Significantly faster at test time than previous methods

Improves anomaly detection accuracy with adversarially learned features

Abstract

Anomaly detection is a significant and hence well-studied problem. However, developing effective anomaly detection methods for complex and high-dimensional data remains a challenge. As Generative Adversarial Networks (GANs) are able to model the complex high-dimensional distributions of real-world data, they offer a promising approach to address this challenge. In this work, we propose an anomaly detection method, Adversarially Learned Anomaly Detection (ALAD) based on bi-directional GANs, that derives adversarially learned features for the anomaly detection task. ALAD then uses reconstruction errors based on these adversarially learned features to determine if a data sample is anomalous. ALAD builds on recent advances to ensure data-space and latent-space cycle-consistencies and stabilize GAN training, which results in significantly improved anomaly detection performance. ALAD achieves state-of-the-art performance on a range of image and tabular datasets while being several hundred-fold faster at test time than the only published GAN-based method.